Re: Anybody using the Progress DataServer for Oracle?

From: Bob O'Connor <oconnorb_at_ix.netcom.com>
Date: 1996/09/25
Message-ID: <3249EFC5.1094_at_ix.netcom.com>#1/1


Gary Elmes wrote:
>
> Is there anybody out there using the Progress DataServer for Oracle?
>
> Is it a good solution for implementing a Progress application in an
> Oracle environment?
>
> Any "traps for young players", "gotcha's" or other inconveniences
> that make the use of the DataServer more of a pain than it is worth?
>
> I've read the manual, so I know about the "official" compromises; it's
> the hidden pitfalls that I'm worried about.
>
> Any feedback much appreciated. Mail or post here.
>
> Rgds,

We've just recently completed a successful install of a complex app using the oracle dataserver. Using 7.3D under DEC Unix on Alpha. It has been pretty painless except for the following:

  1. many instances of VARCHAR too small. The protoora utility creates oracle VARCHARs with the same size as the Progress display format. Sometimes (unexpectedly) we'd try to put more than fits in the field. This blows the session.
  2. you do need an oracle DBA type to deal with the fine points of tuning and sqldba stuff.
  3. while the manual does not make it clear, we used multiple self service clients accessing the schema holder read only (-RO) instead of a multi-user server. In hindsight it may have been better to user the broker/server method.
  4. it is difficult to monitor activity without promon (for lock monitoring). If all clients connect to oracle with the same user id, the sqldba-monitor-locks is of little use also. We're still wrestling with this one.
  5. we found that certain complex progress queries could crash the oracle shadow process. We had to rewrite at least 3 of these. It was faster to do this than to rely on oracle tech support for an answer.

This system has a 1 GB oracle tablespace for all user tables and a separate system tablespace. We chose not to create separate tablespaces for indexes, although we probably should have. Performance has been good, but less than native progress.

Bob Received on Wed Sep 25 1996 - 00:00:00 CEST

Original text of this message